By Nidhi Sinha, Solutions Specialist, Microsoft; Prakash Lekkala, Cloud Solution Architect, Microsoft; and Ramesh Sivasubramanian, VP, Tech & Innovation, Ramco Systems


Ramco Systems is a fast-growing enterprise software player offering multi-tenant cloud and mobile-based software around Global Payroll and HR, ERP, Logistics and M&E MRO for aviation sector.

The company caters to over 500 large enterprises and multinational corporations. One of Ramco’s offerings is a fast and agile attendance system. With Covid-19 outbreak and mandated facemask use in many countries, the company began to look for alternatives to its facial recognition-based attendance system. It partnered with Microsoft and came up with a voice-based verification system. This also alleviated rising privacy concerns with facial recognition-based services.

In order to facilitate a real time attendance system with no delays, the company had to ensure instant user validation (<2 seconds) for thousands of users. It also had to overcome existing speech identification list limit to support 50 user profiles with short user registration process (<15 seconds) and maintain employee privacy throughout the process.

How does the solution work?

The company used Speech to Text (STT) service offered by Microsoft Azure Cognitive Services to come up with a speaker verification system instead of speaker identification, which transformed the registration process. The new solution used 140 voice features to match the registered profile and supported multiple voice diction and tones, including those not used during registration.

• The voice-based system made use of employee ID (short unique codes) to reduce validation time to 2 seconds, resulting in quick onboarding of users.
• Use of STT to identify users, voice profile to verify users, thus overcoming limit of 50 users

Step 1: Thermal sensor detects human presence
Step 2: User dictates the employee ID in front of a microphone
Step 3: Audio sample is sent to Speech to Text service to get employee ID
Step 4: App records user’s audio and buffers it
Step 5: Employee voice profile is loaded, and audio sample is sent to “Speech Recognition” service for verification
Step 6: Attendance system is updated upon successful verification

Benefits of the solution

  • Alleviated employee privacy concerns around facial recognition systems by giving companies complete control over data stored in a secure system.
  • Safe and secure employee return to workplace amid rising Covid-19 fears and strict safety protocols.
  • Accurately identification of speakers, effectively filtering ambient noise.

“Ramco has constantly innovated its products to suit customers’ requirements. With the rising concerns about data privacy in facial recognition-based solutions and the need to devise a solution for markets like Middle East where the traditional facial recognition solutions fall short in user identification, Azure Voice biometric solution has proven to be a secure, robust alternative. Since the solution is scalable, we are focusing on integrating it with other product features and enhancing employee experience. I am sure this solution can completely transform the way we interact with our enterprise systems,” says Ramesh Sivasubramanian, VP, Tech & Innovation, Ramco Systems.

While this has been the use case in practice, there are several other innovations in the pipeline with Ramco Innovations Lab, Singapore, which aim to leverage voice-based verification technology and improve end-user experience. Some of the active implementations are:

1. Enhanced interaction experience with mobile applications with functionality like login screen authentication and navigation through voice-based instructions within the Ramco Unify mobile app

2. Ease of interaction with systems, automated through voice with functions like reaching out to customer support, raising a support ticket, etc.

3. Multi-user device authentication, aiding in role-based access for shared systems

4. Improved employee experience through voice-based attendance, leave requests, etc. with a secure voice verification algorithm.